Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Rudolph]

I just found that EAB 2.10.0 was released.

I went across it, converted images, verified images, converted a font, put that font in a flash container, packed that flash container with the asset compressor, converted the resulting .zopfli file with the BIN2C tool.
And this went very smoothly. :-)

I found one small issue though.
When converting fonts, the "Input Character:" file selection gets cleared every time a new font is selected.

Edit: just found 2.10.1 and the bug is still in while this post is still missing approval.
Edit2: and I just found 2.10.2 which is dated 2023-11-10.

[BRT Community]

Hello Rudolph,

Thank you for your post!

I have been looking into this and the behaviour you are noting is apparent, though I believe this may be related to EAB calculating the applicable Unicode blocks for a given input font, and as such is an indication that you may want to review the input character set as the new font might not these characters.

In any case I will liaise with the development team on this point to see if we can make some updates for the next EAB revision.

Best Regards,
BRT Community

[Rudolph]

I have been looking into this and the behaviour you are noting is apparent, though I believe this may be related to EAB calculating the applicable Unicode blocks for a given input font, and as such is an indication that you may want to review the input character set as the new font might not these characters.

Well, the "Input Character" file is there to limit the glyphs of a font to the desired subset, not to wish glyphs into existing. :-)
So I would say that is more likely that someone needs to convert a bunch of different fonts to the same subset of glyphs as it is
that the different fonts need to be converted for different sets of glyphs.
And even when someone needs to change the "Input Character" file for every font, this does not require any more clicks than with the current implementation while everyone who does not need to change the "Input Character" file must select the same file again and again.

Even only browsing thru fonts to find one that has all the glyphs needed for the selected set or looks ok for the selected set is
a tiny bit more cumbersome now.

And yeah, I am having fun here "complaining" over this first-world useability issue as I have not found anything more serious so far - good job! :-)

[BRT Community]

Hello,

Thank you for your input.

There has been an update from the development team on this point, and they are going to make a change in the next version of EAB which should prevent the input character file from resetting if the input font has changed.

Best Regards,
BRT Community

[Rudolph]

Well now, EAB 2.11.0 is out and it clears the Input Character file when you change fonts.
L8 is still selected when going for extended format, "Reserve Space" is still not selected by default and ASTC 4x4 is still the default.

And it still generates the 7.4MB "SampleApp" everytime.
At least there is a way to stop this, rename C:\Users\Public\Documents\EVE Asset Builder\tools\SampleApp

[BRT Community]

Hi Rudolph,

Yes, 2.11.0 is now out,

We were not able to make some other changes in time for this one but rest assured your suggestions are with the R&D team for future consideration,

Best Regards, BRT Community

[Rudolph]

I converted some fonts with EAB 2.12.2 and it does not reset anything anymore when swithing the font - and I really like that. :-)

However, there is an issue in the flash utilities for a while now, the generated .map file looks like this now:
default-fl.blob                     : 0     : 4096
Rubik-Bold_24_ASTC.glyph            : 4096  : 34368
Rubik-Bold_24_ASTC.xfont            : 38464 : 309 
Rubik-Bold_24_ASTC.xfont.padding    : 38773 : 11   
Rubik-Regular_20_ASTC.xfont         : 38784 : 309 
Rubik-Regular_20_ASTC.xfont.padding : 39093 : 11   
Rubik-Regular_20_ASTC.glyph         : 39104 : 34368
Rubik-Regular_24_ASTC.xfont         : 73472 : 309 
Rubik-Regular_24_ASTC.xfont.padding : 73781 : 11   
Rubik-Regular_24_ASTC.glyph         : 73792 : 34368

The issue is this .padding "file" that was introduced a while ago.
This does not only add a bunch of mostly useless lines, this had the effect that the size for the .xfont file can no longer be directly
used for the CMD_FLASHREAD call as CMD_FLASHREAD requires the number of bytes to be read to be a multiple of 4.
In a previous version of EAB the length of the .xfont file was set to 320 and while that "wasted" 11 bytes of precious RAM_G, at least it worked.

Can we please get rid of the .padding entry again and perhaps round up the length of the file to the next multiple of 4?
That would be 312 in this case.
Yes, that still is 3 bytes more than necessary. But you need to use 312 as a minumum with CMD_FLASHREAD anyways and the destination address for CMD_FLASHREAD needs to 4-byte aligned as well, so you can not place the .xfont files of 309 bytes length back-to-back anyways.
I rather have small holes in the .map and not extra lines that are not really helping.